The Pitfalls of UM

Unified messaging (UM), a potentially high-growth area, has yet to really take off. Carriers blame this on a variety of things, including poor text-to-voice abilities and a need for standards. The wireless e-mail-fax-voice-mail service's lackluster performance is sending technicians back to the drawing board to fine-tune software and alleviate problems.

Robert Wohnoutka, Lucent product manager of UM solutions, said several issues must be addressed before UM attracts a strong market. One of his top concerns is text-to-voice capability. Sometimes called the "drunken Swede" because its synthetic voice is hard to understand, UM's problem stems from software reading each character rather than sets of characters. According to Wohnoutka, Lucent's latest application will address the text-to-voice problem head-on. Instead of taking each letter alone, the new software is given a database of phonemes, or character sets, such as "ch," "ough" or "ea." The new software will make the voice easier to understand and should speed translation in the process. UM's voice never will be perfect, Wohnoutka said, but it can be better.

One area where the voice could improve is in reading numbers. Because there are only 10 number characters, the software can be programmed to speak each one perfectly. Lucent's software technicians are working on making that possible. But Lucent developers have admitted that it has ignored the reverse -- voice-to-text.

With so many dialects, new slang terms and different voice tones, voice-to-text software needs to be particularly sensitive. The problem is that software must be trained on each user's voice and speech in order to convert speech easily. Most users don't want to spend time doing this, so software is left to do the extra work.

Voice-recognition, word-context and dictionary applications must be upgraded to make voice-to-text work better. For now, technicians are satisfied with industry-wide voice-to-text capabilities, which are nearly 95% accurate.

Greg Baltzer, Dialogic vice president of public network marketing, said his latest voice-to-text evaluations found it to be surprisingly good.

"That's one of those technologies that they are going to refine and refine forever," Baltzer said.

What makes enhancing UM even more difficult is the environment in which users typically want to use the voice-to-text feature. Commonly, users want to dictate messages while they are driving, a difficult environment for voice-to-text software to work. Due to its sensitive algorithms, the software hears everything, including background noise. Differentiating and tuning out excess noise is difficult.

Mark Ozur, Unisys PulsePoint Communications president, said one solution is for software to tell the computer to hear only what is important. Speech recognition always will be UM's greatest challenge, he added.

STANDARDS Other challenges facing software developers are not as long term. Application and handset standards already are taking shape.

Many UM users want to get their messages while on the road. That typically means integrating multiple systems, which is not an easy task. But standards are helping. IMAP 4, a still-developing Internet messaging protocol, allows clients to access their messages, no matter what software their carriers use. In the near future, all message stores in the world will have an IMAP 4 interface.

More WAP handset standards also will make UM use easier. Users want large features, but smaller phones. Developers say WAP handsets need to be standard so UM functions won't be limited to voice-centered features.

WAP handsets have some drawbacks of their own. They only use server-side UM technology and traditionally have poor battery life. Developers expect upgrades to address these problems, however.

With standard implementation just around the corner, carriers offering UM must decide whether to move quickly and just replace their whole systems or upgrade piece-by-piece until they get there. Dialogic's Baltzer said carriers should replace their entire systems.

For instance, many carriers use the common call-answering platform Octel Sierra. Using it alone can't offer UM, and building on that platform will be difficult. Using an open system solution will be easier. Having a database that works with e-mail or a common software-oriented application such as those found in data solutions is always easier than building on a platform that can't integrate quickly.

"Without standards, the systems have a fundamental problem," he said. "In the long run, upgrading will not work."

After dealing with the huge cost of system replacement, carriers have become cautious. Many are researching the UM market a second or third time before making a decision.

A spokesperson for one national carrier that is on the verge of offering UM said it would monitor the market-pull before publicly announcing its intention to offer UM. Analysts, however, said UM is the next service that will give carriers the opportunity to break away from the pack and differentiate their services.

INTEGRATION Gary Hermansen, Glenayre Electronics senior vice president & general manager of the consumer products group, disagrees. UM alone probably won't make carriers unique, he said. UM's defect is not so much a software problem as it is an integration problem. UM lacks integration with a personal information manager (PIM).

For out-of-the-office executives, scheduling appointments and making contact is essential. Retrieving and sending messages, whether fax, voice or e-mail, will do little good without a database to store the information in a useable context.

Say, for example, a user is on the road, and he has an e-mail requesting an appointment next Thursday at 2 p.m. Is he going to pull over, find his calendar and a pen, write down the appointment, then respond to the e-mail? A PIM attached to a wireless phone would allow him to enter the information once and have it serve two purposes: respond to the message, and schedule and keep track of information given through the message. Once WAP standards are implemented, Hermansen said, PIM and UM integration will drive the technology.

Probably more important for WAP standards is the need for higher bandwidth for all of this data. Without it, UM is sure to fizzle.

"Here (users) want us to push so much data their way," Ozur said. "But they don't want WAP phones because they are larger than they are used to."

Handsets such as the Nokia 9000 and Qualcomm pdQ phone, two previously under-appreciated products, have a new market. According to Ozur, they are the future of UM.

END-USER CONTROLS Other software upgrades needed for UM to take off will include more end-user controls, such as filtering capabilities.

Customers want their phones to tell them which messages are important and which can wait. Developers say UM users already have proved they don't shy away from reading handsets as standard users often do, but they also agree that users want to control how much they read.

Fast scrolling, voice commands and quick-search features will make the difference. Being able to quickly ascertain which reply goes with which message also will help.

Glenayre's Hermansen uses UM himself and said one problem he has is remembering if the e-mail he is reading is a reply to a voice mail he sent or a fax. With more choices comes more confusion. Software will need to eliminate some of that confusion.

Some users will seek even more control. Some will want true UM with a single directory. For example, if a user deleted a message from his handset, it will be gone from all databases. Others only want integration of the handset or the ability to retrieve voice messages, e-mail and faxes from their phones. With integrated handsets, if a message deletes, it will disappear from an intermediary server but not necessarily the original source.

UM's future will weigh heavily on software developers' ability to address their products' failures and add special features. But, they say they are ready to take the challenge.

IN THE FUTURE PulsePoint's Ozur said e-mail vendors will drive UM because they already have e-mail components and are able to integrate multiple clients. But even software giants such as Microsoft will need some help, and they will look to companies such as PulsePoint for that help.

Developers at Lucent said UM's future depends on its ability to improve personal productivity. UM must add enough individual value to wireless service for it to move forward. Most importantly, UM will have to prove cost savings.

Currently, UM's biggest problem is its voice-centered capabilities, which are only adequate, but getting better all the time. Added features such as Internet content and integrated PIM are on their way to making UM even more effective. With all these solutions, developers are confident UM will be well worth the cost.

"I don't think there is any doubt that (UM) will be worth it because it is just a better way of doing things," said Lucent's Wohnoutka.

Information is the name of the game, and the winner is he who can provide the most cost effective, interactive, simple way for subscribers to access their important information. PCS '99 was a haven for data solutions of every ilk, from FusionOne's Internet synchronization technology to Phone.com's mobile portal platform to SmartServ Online's data store service center. In the mix were several companies that have chosen to focus on instant messaging (IM). These companies have extended IM's scope, once solely the domain of the wireline world, to the wireless world.

Tegic Communications plans to capitalize on the PC chat revolution by adding an IM application to a version of its T9 text input software. The technology, which is compatible with Yahoo, Microsoft and America Online's PC-based IM platforms, allows users to see whether others are on-line and send instant text messages to people on their buddy lists, which are updated automatically. The company plans to deliver the application in three forms: as an embedded application bundled with T9; as an application for SIM toolkits; and in WAP phone-compatible form.

However, @mobile.com's plan is to target the existing 28 million handsets in the market. It offers a service-bureau solution, so @mobile's gateway server connects the Internet to the wireless network. Messages to the phone are sent via short message service, but because 2-way SMS virtually is unavailable, customers can request stock information, news, etc., through a series of keystrokes that correspond to predefined messages. This process is much the way teenagers use short codes to send messages back and forth to each other, said Michael Buhrman, @mobile.com president & CEO. The technology uses SMS to deliver messages, so there is no additional infrastructure cost for carriers and no handset modification.

"The carrier gets to enter the 2-way marketplace in an inexpensive way and learns what its customers' preferences are," Buhrman said. "In terms of building applications for browser phones, this is a great way for them to characterize their customers, find out what they like and don't like."

Marrying the simple and the complex, HiddenMind Technology has taken the IM idea a few steps further. Although users can use the system to chat, the company's leaders envision its technology having a broader appeal. They see it as a means to connect and streamline business processes.

Although the company often is compared with Phone.com, Chairman & CTO Kristopher Tyra explained that one difference is that HiddenMind uses push technology, and Phone.com uses pull technology. On a basic level, the company joins IM with wireless, but its higher-end products simplify workflow communications. For example, HomeWrite has licensed HiddenMessenger for its homebuilders and homeowners Web portal. The system allows homebuilders, subcontractors, suppliers and homeowners to be on the same page, so to speak. Each group can track home construction projects and communicate with one another.

Nortel Networks was one of HiddenMind's early projects, and it has struck deals with Omnipoint Technologies and BellSouth Mobility DCS.